Tipping Points and Early Warning Signals in the Climate-Carbon System
Clarke, J
Date: 8 January 2024
Thesis or dissertation
Publisher
University of Exeter
Degree Title
Doctor of Philosophy in Mathematics
Abstract
This is a thesis about tipping points and early warning signals. The tipping points investigated
are related to various components of the climate-carbon system. In contrast, the work on
early warning signals has more generic applications, however in this thesis they are analysed
in the context of the climate-carbon system. The thesis ...
This is a thesis about tipping points and early warning signals. The tipping points investigated
are related to various components of the climate-carbon system. In contrast, the work on
early warning signals has more generic applications, however in this thesis they are analysed
in the context of the climate-carbon system. The thesis begins with an introduction to the
climate-carbon system as well as a discussion of tipping points in the Earth system. Then a more
mathematical summary of tipping points and early warning signals is given. An investigation
into the ‘compost bomb’ is undertaken, in which the spatial structure of soils is accounted
for. It is found that a hot summer could cause a compost bomb. The effect of biogeochemical
heating on the stability of the global carbon cycle is investigated and it is found to play only
a small role. The potential for instabilities in the climate-carbon cycle is further investigated
when the dynamic behaviour of the ocean carbon cycle is accounted for. It is found that some
CMIP6 models may be close to having an unstable carbon cycle. Spatial early warning signals
are investigated in the context of more rapidly forced systems. It is found that spatial early
warning signals perform better when the system is rapidly forced compared with time series
based early warning signals. The typical assumptions about white noise made when using
early warning signals are also studied. It is found that time correlated noise may mask the early
warning signal. It is shown that a spectral analysis can avoid this problem.
Doctoral Theses
Doctoral College
Item views 0
Full item downloads 0